HIVE Digital Technologies signs five-year GPU cloud deal valued at about $350 million

HIVE Digital Technologies Ltd announced that it has signed a five-year GPU cloud services agreement with an unnamed, investment-grade enterprise customer valued at approximately $350 million. The company characterized the deal as focused on providing GPU cloud services, a category of computing that uses graphics processing units to accelerate AI training and inference, analytics, and other compute-heavy workloads.

The agreement length, five years, is designed to give HIVE longer visibility into contracted demand. In its statement, HIVE did not name the customer or disclose the precise timing of service commencement, the geographical scope of delivery, or the specific service levels included in the contract.

The company also did not provide a breakdown of how the roughly $350 million valuation translates into monthly or yearly revenue, nor did it share any unit economics such as pricing per GPU, utilization assumptions, or expected capacity build-out tied to the customer’s needs. Those details matter for understanding how much of the contract will flow through as gross margin versus infrastructure costs.

Because the customer is described only as “investment-grade” and not identified, investors have limited ability to benchmark the deal’s risk profile or compare it with other enterprise compute arrangements. HIVE did not provide information about whether the contract includes minimum volume commitments, take-or-pay provisions, termination rights, or performance credits for service levels.

In broader terms, GPU cloud offerings have gained traction as enterprises increasingly move AI workloads into managed infrastructure rather than buying and operating hardware directly. Unlike general-purpose cloud compute, GPU-accelerated environments are often required for running modern machine learning models and for speeding up data processing pipelines, which can make contracted access more appealing to organizations that need predictable capacity.

For a company like HIVE, which positions itself around providing GPU-enabled computing, longer-duration agreements can support more stable planning for data center operations and hardware purchasing cycles. GPU infrastructure typically involves significant upfront investment, including data center power and cooling, and companies often seek multi-year contracts to reduce revenue volatility.

What HIVE did not disclose in the announcement is also important. The statement did not specify what GPU platforms or software stack the customer will use, whether the arrangement includes dedicated or shared GPU capacity, or how the company will handle future scaling if the customer expands workloads during the five-year period.

Going forward, investors will likely watch for additional disclosure around how much of the contract is expected to be recognized in the coming quarters, whether HIVE provides further operational detail about the hosting environment, and whether it expands the customer relationship into additional services beyond the initial GPU cloud scope.